ADDICTION, SELF-SIGNALLING, AND THE DEEP SELF 1 RICHARD HOLTON . INTRODUCTION It is easy to see addiction as destroying intentional action altogether; easy to see addicts as swept away by a welter of compulsive behaviors. But the empirical evidence has increasingly indicated that such a picture is not right. 2 While addiction does involve plenty of involuntary states—most centrally, cue-driven cravings—these do not in themselves amount to behaviour. Behaviour comes from the interaction of these involuntary states with more familiar deliberative processes. So in trying to understand why addicts act as they do, we need to understand these processes, and the beliefs, desires, intentions, and other mental states that are involved. In this context then, it will matter whether addicts think that it is worth giving up, and whether they think that they are able to do so. What information will they have? hey will doubtless have been told many things. But they will also be observers of their own behaviour, hypothesizing about their situation and capacities on the basis of what they have done. Here they need not just be passive self-observers. hey may be acting partly in order to gain relevant information; that is, they may be self-signalling. Forthcoming in Mind and Language 1 Versions of this paper were presented at the Addiction, Self and Self-Knowledge workshop run by Mind and Language in London, March ; and at the Self-deception, Self-signaling, and Self-control workshop at the IAST Toulouse, June . Many thanks to the audiences there, and to Rae Langton, Adam Bates and the referees for this journal, for comments and discussion. Special thanks to Drazen Prelec for first making me think about self- signalling. 2 he evidence for this is now very strong and can be found in many writings. Kent Berridge and I reviewed some of it in an earlier piece (Holton and Berridge ), but for a fuller and more recent summary, see (Pickard, ).